When done well, systematic reviews and meta-analyses are regarded as the gold standard for summarising the state of research on a given topic. They provide a solid foundation for evidence-based decisions in academia and education policy. However, producing such reviews is labour-intensive and usually involves a huge time commitment. One of the best-known meta-analyses in educational research, John Hattie’s ‘Visible Learning’, summarises over 800 meta-analyses on factors influencing learning success and took a full 15 years to complete. In fast-moving fields of research such as educational research, this is a problem, particularly when urgent questions, such as those arising during the COVID-19 pandemic, require rapid answers.

This is where the study by researchers from the Hector Institute and other scientists comes in. They demonstrate how Artificial Intelligence (AI) can make the production of reviews and meta-analyses more efficient. AI tools can assist with literature reviews through automation processes, extract data and support analyses. This saves time and costs and reduces human biases that can arise in decision-making. Nevertheless, such tools have hardly been used in educational research to date. It is likely that such AI tools are not yet sufficiently well-known and that there are concerns about their complexity.

To lower the barriers to using AI tools, the authors reviewed 282 available tools and selected seven that are particularly suitable: transparent, easily accessible and usable without programming knowledge. ASReview is also among the selected tools. A tutorial paper, to which Dr Tim Fütterer from the Hector Institute for Empirical Educational Research has also contributed, demonstrates in detail how this tool can be used in reviews.

These AI tools can accelerate research, but only if they are combined with good scientific practice. Transparency is crucial here, both in the documentation of search strategies and in the disclosure of the algorithms used. Furthermore, human oversight remains indispensable. All seven AI tools leave the final decisions to the researchers, which is important for quality and scientific accuracy.

Furthermore, the question of the sustainability of such AI tools arises. Many freely accessible AI tools suffer from a lack of maintenance. One application encountered technical difficulties, whilst another faced access issues due to security standards. The authors therefore call for long-term funding models to ensure the stability and accessibility of AI tools that are open source and have not been developed for commercial purposes. Otherwise, there is a risk of greater dependence on commercial providers.

The study makes it clear: AI is not a substitute for scientific rigour, but a complement that paves the way for more efficient and inclusive research processes. Prerequisites include transparency, human oversight and sustainable structures for the further development of AI tools.

Publication 
Fütterer, T., Campos, D. G., Gfrörer, T., Lavelle-Hill, R., Murayama, K. & Scherer, R. (2025). AI tools for systematic literature reviews and meta-analyses in educational psychology: An overview and a practical guide. Learning And Individual Differences, 126, 102849. 
https://doi.org/10.1016/j.lindif.2025.102849

ASReview-Tutorial 
https://doi.org/10.1177/25152459261442150